Navigating AI Usage and Outcomes-Based Pricing Decisions

Pricing has never been easy, but identifying the right pricing strategies for AI features and outcomes-based pricing poses new challenges for SaaS companies.Traditionally, SaaS providers have had a relatively fixed set of criteria to base their pricing on: Seats. Number of transactions. Record bands.The challenge posed by AI features and outcome-based pricing is a broader set of unknowns that SaaS providers must account for.

AI Feature Usage and Cost Exposure

The key question in pricing a new AI feature is how much customers will actually use it. Most SaaS providers license an AI engine from an AI provider, such as OpenAI or Anthropic, and these providers tend to charge based on usage volume. In most cases, SaaS companies cannot, or prefer not to, charge a usage-based fee for new AI features, so they must estimate how much the average customer will use and set the price accordingly.

This creates a balancing act. If they charge too much for the new feature, it will hinder adoption, which in turn could diminish the solution’s value to customers. On the flip side, if they don’t charge enough, or anything at all, and customer-use is high, it becomes a quiet margin killer.

Defining Outcomes When Customers Control Results

While outcomes, or value-based pricing, differ structurally from AI feature pricing, the risk is similar: The customer creates the unknown. Outcomes-based pricing shifts the pricing metric from a clearly defined criterion (e.g., seats, records) to a performance variable. Was the customer able to resolve more cases? Close more deals? Raise more money?

The basic issue is that, regardless of how robust or powerful a SaaS platform is, much of the “success” of the solution is out of the provider’s control. Outcomes depend on a range of factors, including proper implementation, the data fed into the platform, integration with other systems, the customer’s processes, training, and the effectiveness of their change management approach. Who or what deserves credit for an improvement is also subject to the human tendency for customers to credit their own efforts for their successes and to blame their tools for their failures. This makes it challenging to define a “fair” outcome.

Fundamentals of Pricing Research Still Apply

Although these are new developments in pricing strategy, the core principles of pricing research still apply. That means the question isn’t whether pricing research still works, but how to adapt it to account for behavior that hasn’t happened yet. Pricing research typically focuses on what customers or prospects pay today, what they consider acceptable pain points, and the perceived value of the job-to-be-done, using methods that elicit pricing elasticity, trade-offs, and thresholds.

These dimensions remain important in AI and value-based pricing research because they anchor willingness to pay for a solution. However, what’s missing is whether customers will use a new feature and, if they do, to what extent. In theory, it would be as simple as asking customers to estimate how much they would use a new feature and what benefits they would expect.

In reality, buyers are better at describing what they do today and why than they are at predicting their future behavior. This is especially true when a feature or solution involves changing an existing process. Therefore, the research objective is to sufficiently profile customer behavior to provide a rational basis for anticipating their usage, in terms of volume or outcomes.

Using Customer Context to Inform Pricing Decisions

Market insights can provide framing for pricing and packaging strategies. The first step is for SaaS marketing teams to develop hypotheses about the threshold conditions required for customers to adopt the solution/feature and the expected improvements.

Illustrative areas to explore include:

  • How do they do the job or task today, and how satisfied are they with the outcomes?
  • Benchmarking Data: How much time do they spend on the activity? How much data do they have? How many customer records? How many transactions? What’s the quality of that data?
  • Do they have the teams, skillsets, governance/risk posture, and mindsets in place to adopt change?
  • What are their beliefs about specific outcomes a new solution could produce? What do they view as credible vs. marketing math?

Gathering data from customers and prospects can help validate and refine those assumptions.

Establishing these market parameters enables SaaS marketers and pricing teams to understand the use-case context. For example, can 10% or 50% of an account’s customer support calls be handled by a virtual agent? The answer impacts volume of use and perceived value.

These insights can be gathered from customers or prospects, depending on the solution and its primary target. For example, new features are typically aimed at customers; thus, customers could provide the necessary insights. Outcome-based pricing, on the other hand, is a strategy for winning new business and would require gathering insights from prospects.

Quantitative and qualitative insights can help inform the decision. Qualitative approaches can be especially helpful at the early stages of pricing initiatives. Surveys can provide quantitative data. Qualitative can provide directional information that is often as useful, especially at the early stages of pricing exercises or when the solution or its outcome is complex.

Informed Pricing Strategies

Pricing research is about mitigating risk and developing a systematic rationale for pricing decisions. Even in the best circumstances, pricing strategies are informed estimates of the optimal price for the market and the SaaS provider and must be revisited over time. However, developing a deeper, more systematic understanding of how the market thinks and acts can guide pricing strategies and reduce the risks of getting it wrong. It also provides a shared foundation for pricing decisions across the product, sales, marketing, and finance functions.

If you’d like to read more about Isurus’ perspectives on pricing research, check out these related posts:

How to identify the right price point in B2B — A practical walkthrough of combining current spend, buyer context, willingness to pay, and price acceptability to arrive at a pricing strategy that reflect real buying conditions.

Beyond willingness to pay — Why WTP data alone often overstates demand—and how budget ownership, internal priorities, and organizational constraints shape buying decisions.

Alternatives to conjoint — When simpler pricing research approaches often provide clearer, more actionable guidance than complex models.

Anchoring in pricing research — How buyers’ existing spend, tools, and expectations influence their reactions to price, and why understanding these anchors is critical for interpreting pricing feedback.

For another perspective, check out the following articles:

If you’d like to talk about a pricing challenge you are facing, you can reach us via our contact page: https://isurusmrc.com/contact